On Tue, Aug 26, 2003 at 09:35:59AM -0500, David Anders wrote:
> Greg,
> 
> turns out its a simple fix, on the pl2303 you just need to set OPOST to 
> off on in the set_termios function. this prevents other apps from making 
> a change to the status. after making the addition, echo, cat, and my 
> test app work fine without lockups. this still leaves the issue of not 
> having onlcr translation, but this can be done in userland with a fifo 
> and a filter. this fix should probably be implemented in all of the 
> usb-rs232 adapters since the way the usb to rs232 interface is created 
> none of these devices should support opost. in addition this problem 
> probably also exists with usb modem devices such as the acm module. i'm 
> currently looking at the multitech usb modem using hylafax, so i should 
> know something shortly. attached is a diff of the pl2303.c change on 
> kernel 2.4.21 with your last posted patch, and a diff for 2.6.0-test3 
> with no patches. please have a look and let me know.

No, this is just papering over the real problem and is not a solution,
sorry.

How about a tweak to the driver to never fail a write.  It's pretty easy
in 2.6 to make this change if you want to test out if this solves your
problem (but from the logs I don't think it will...)

thanks,

greg k-h


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to